Authors: Atul Gawande
Ross led me into his office. Soft-spoken and unhurried, he has a soothing demeanor that fits perfectly with his line of work. Quinlan’s kind of problem, he told me, is the one he sees most frequently. Chronic back pain is now second only to the common cold as a cause of lost work time, and it accounts for some 40 percent of workers’ compensation payments. In fact, there is a virtual epidemic of back pain in this country today, and nobody can explain why. By convention, we think of it as a mechanical problem, the result of
misplaced stress on the spine. We therefore have had some sixty years of workplace programs, and now there are even “back schools,” which teach the “correct way to lift,” among other things. Despite the fact that the number of people who engage in manual labor has steadily declined, however, more people have chronic back pain than have ever had it before.
The mechanical explanation is almost certainly wrong, Ross noted. It’s true that lifting something the wrong way can cause a muscle pull or a slipped disk. But that sort of strain occurs in almost everyone at some time, and in most people it never becomes a persistent problem. Scores of studies have looked for physical factors that can predict which acute back injuries will evolve into chronic back pain, but they haven’t found any. For instance, doctors used to assume that damaged disks were associated with pain, but recent findings have not borne this out. Spinal MRI scans show that most people without back pain have disk bulges. Conversely, a large percentage of patients with chronic back pain, like Quinlan, are found to have no structural lesion. And even among those with abnormalities there is no relation between the severity of the pain and the severity of the abnormalities.
If the condition of your back doesn’t predict whether you’ll get chronic back pain, what does? Well, it’s the mundane stuff that neither doctors nor patients much like to consider. Studies point to such “inorganic” factors as loneliness, involvement in litigation, receipt of workers’ compensation, and job dissatisfaction. Consider, for example, the epidemic of back pain in the medical profession itself. Disability insurers once saw doctors as ideal customers. Nothing stopped doctors from working—not years of stooping over operating tables, not arthritis, not even old age. Insurers used to try to outbid one another with cheap rates and generous benefits to attract their business. In the last few years, however, the number of doctors with disabling back or neck pain has risen dramatically. Needless to say, doctors aren’t suddenly being required to carry heavy packages around. But
one known risk factor has been identified: with the growing role of managed care, job satisfaction in the medical profession has plunged.
The explanation of pain that has dominated much of medical history originated with Rene Descartes, more than three centuries ago. Descartes proposed that pain is a purely physical phenomenon—that tissue injury stimulates specific nerves that transmit an impulse to the brain, causing the mind to perceive pain. The phenomenon, he said, is like pulling on a rope to ring a bell in the brain. It is hard to overstate how ingrained this account has become. Twentieth-century research on pain has been devoted largely to the search for and discovery of pain-specific nerve fibers (now named A-delta and C fibers) and pathways. In everyday medicine, doctors see pain in Cartesian terms—as a physical process, a sign of tissue injury. We look for a ruptured disk, a fracture, an infection, or a tumor, and we try to fix what’s wrong.
The limitations of this mechanistic explanation, however, have been apparent for some time. During the Second World War, for example, Lieutenant Colonel Henry K. Beecher conducted a classic study of men with serious battlefield injuries. In the Cartesian view, the degree of injury ought to determine the degree of pain, rather like a dial controlling volume. Yet 58 percent of the men—men with compound fractures, gunshot wounds, torn limbs—reported only slight pain or no pain at all. Just 27 percent of the men felt enough pain to request pain medication, although such wounds routinely require narcotics in civilians. Clearly, something that was going on in their minds—Beecher thought they were overjoyed to have escaped alive from the battlefield—counteracted the signals sent by their injuries. Pain was becoming recognized as far more complex than a one-way transmission from injury to “ouch.”
In 1965, the Canadian psychologist Ronald Melzack and the British physiologist Patrick Wall proposed that the Cartesian model be replaced with what they called the Gate-Control Theory of Pain.
Melzack and Wall argued that before pain signals reach the brain they must first go through a gating mechanism in the spinal cord, which could ratchet them up or down. In some cases, this hypothetical gate could simply stop pain impulses from getting to the brain. In fact, researchers soon identified a gate for pain in a portion of the spinal cord called the dorsal horn. The theory explained such ordinary puzzles as why rubbing a painful foot makes it feel better. (The rubbing sends signals to the dorsal horn that close the gate to nearby pain impulses.)
Melzack and Wall’s most startling suggestion was that what controlled the gate was not just signals from sensory nerves but also emotions and other “output” from the brain. They were saying that pulling on the rope need not make the bell ring. The bell itself—the mind—could stop it. Their theory prompted a great deal of research into how factors such as mood, gender, and beliefs influence the experience of pain. In one study, for example, researchers measured pain threshold and tolerance levels in fifty-two dancers from a British ballet company and fifty-three university students using a standard method called the cold-pressor test. The test is ingeniously simple. (I tried it at home myself.) After immersing your hand in body-temperature water for two minutes to establish a baseline condition, you dunk your hand in a bowl of ice water and start a clock running. You mark the time when it begins to hurt: that is your pain threshold. Then you mark the time when it hurts too much to keep your hand in the water: that is your pain tolerance. The test is always stopped at a hundred and twenty seconds, to prevent injury.
The results were striking. On average, female students reported pain at sixteen seconds and pulled their hands out of the ice water at thirty-seven seconds. Female dancers went almost three times as long on both counts. Men in both groups had a higher threshold and tolerance for pain—as expected, since studies show women to be more sensitive than men to pain, except during the last few weeks of pregnancy—but the difference between male dancers and male nondancers was nearly as large. What explains the difference?
Probably it has something to do with the psychology of ballet dancers—a group distinguished by self-discipline, physical fitness, and competitiveness, as well as by a high rate of chronic injury. Their driven personalities and competitive culture evidently inure them to pain: that’s why they are able to perform through sprains and stress fractures, and why half of all dancers develop long-term injuries. (Similar to other nondancing males, I started to feel pain at around twenty-five seconds; but I had no trouble keeping my hand in for the whole hundred and twenty seconds. I will let others speculate on what this says about the submissiveness inculcated in surgical residents.)
Other studies along these lines have shown that extroverts have greater pain tolerance than introverts, that drug abusers have low pain tolerance and thresholds, and that, with training, one can diminish one’s sensitivity to pain. There is also striking evidence that very simple kinds of mental suggestion can have powerful effects on pain. In one study of five hundred patients undergoing dental procedures, those who were given a placebo injection and reassured that it would relieve their pain had the least discomfort—not only less than the patients who got a placebo and were told nothing but also less than the patients who got a real anesthetic without any reassuring comment that it would work. Today, it is abundantly evident that the brain is actively involved in the experience of pain, and is no mere bell on a string. Today, every medical textbook teaches the Gate-Control Theory as fact. There’s a problem with it, though. It doesn’t explain people like Roland Scott Quinlan.
Gate-Control Theory accepts Descartes’s view that what you feel as pain is a signal from tissue injury transmitted by nerves to the brain, and it adds the notion that the brain controls a gateway for such an injury signal. But in the case of Quinlan’s chronic back pain, where is the injury? Or take something like phantom-limb pain. After amputation of a limb, most people suffer a period of constant, intractable burning or cramping that feels exactly as if the limb were still there. Without a limb, however, there are no nerve impulses for
the gate to control. So where does the pain come from? The rope and clapper are gone, but the bell can still ring.
One spring day in 1994, Dr. Frederick Lenz, a neurosurgeon at the Johns Hopkins Hospital, brought to his operating table a patient suffering from severe hand tremors. The patient, whom I’ll call Mark Taylor, was only thirty-six, but over the years his hands had come to shake so violently that the simplest of tasks—writing, buttoning his shirt, drinking from a glass, or typing on his keyboard at his job as a purchasing agent—grew absurdly difficult. Medications failed, and he lost jobs more than once because of his difficulties. Desperate for a return to a normal life, he agreed to a delicate procedure: brain surgery that would destroy cells in a small structure called the thalamus, which was already known to contribute to such excessive stimulation of the hands.
Taylor had another big problem, though: for seventeen years, he had struggled with a severe panic disorder. At least once a week, while he was working at his computer terminal or was at home in the kitchen feeding a child, he would suddenly be overcome by severe chest pains, as if he were having a heart attack. His heart would pound, his ears would ring; he would grow short of breath and would have an overwhelming urge to escape. Nevertheless, a psychologist Lenz consulted assured him that the disorder was unlikely to hinder the operation.
Initially, Lenz says, everything went as he had expected. He injected a local anesthetic—the operation is done with the patient awake—and burred a small opening in the top of Taylor’s skull. Then he cautiously inserted a long, thin electrical probe deep inside, right down into the thalamus. Lenz talked to Taylor the whole time, asking him to stick out his tongue, to move a hand, to do any of a dozen other tasks that showed he was all right. The danger in this type of surgery is that it might destroy the wrong cells: the thalamic cells involved in tremor lie just fractions of a millimeter away from cells that are essential for sensation and motor activity. So
before cauterizing with a second, larger probe, the surgeon had to find the right cells by stimulating them with a gentle electric pulse. The probe was in a portion of Taylor’s thalamus that Lenz labeled Site 19, and he zapped it with low voltage. He had been here a thousand times before, and typically, he told me, zapping the site makes people feel a prickle in the forearm. Sure enough, this is what Taylor felt. Lenz then zapped an adjacent area he labeled Site 23, where stimulation generally produces a mild and very ordinary tingling in the chest. This time, however, Taylor felt an unexpectedly far more harsh pain—in fact, the exact chest pain of his panic attacks, along with the suffocation and instant sense of doom that always accompanied them. It made him cry out and nearly leap off the table. When Lenz stopped the stimulation, however, the sensation disappeared, and Taylor became instantly calm again. Puzzled, Lenz zapped Site 23 once more, and found that doing so produced the same effect again. He stopped, apologized to Taylor for the discomfort, and went on to locate the cells controlling his tremor and to cauterize them. The operation was a success.
Yet even as Lenz completed the procedure, his mind was racing. Only once before had he seen anything like this kind of effect. It was in a sixty-nine-year-old woman with a long history of difficult-to-manage anginal pain that came on not only with strenuous activity but even with mild physical exertion that wouldn’t be expected to stress her heart. Performing a similar operation on her, Lenz found that stimulating the microscopic section of her brain that usually triggered mild chest tingling had instead, as with Taylor, brought on her more severe and familiar chest pain—a sensation she described as “deep, frightful, squeezing.” The implications might have easily been lost, but Lenz had spent many years researching pain and realized that he had witnessed an important and telling effect. As he later noted in a report published in the journal
Nature Medicine
, the response in these two patients was wildly out of proportion to the stimulus. What in most people produces no more than a tingle was torture to them. Areas of the brain governing ordinary sensations
appeared to have become abnormally sensitized—set to fire in response to perfectly harmless stimuli. In the woman’s case, her chest pain had begun as a signal of her heart disease but now appeared in circumstances that did not reflect anything like an impending heart attack. Even more oddly, in Taylor’s case, the pain had not begun with any such bodily damage, but with his panic disorder, which is understood to be a psychological condition. Lenz’s findings suggest that, in fact, all pain is “in the head”—and further that sometimes, as with Mark Taylor or perhaps Roland Scott Quinlan, no physical injury of any kind is needed to make the pain system go haywire.
This is the newest theory of pain. Its leading proponent is, once again, Melzack, who abandoned Gate-Control Theory in the late 1980s and began telling incredulous audiences to revise their understanding of pain once again. Given the evidence, he now says, we should stop thinking that pain or any other sensation is a signal passively “felt” in the brain. Yes, injury produces nerve signals that travel through a spinal-cord gate, but it is the brain that generates the pain experience, and it can do so even in the absence of external stimuli. If a mad scientist reduced you to nothing but a brain in a jar, Melzack says, you could still feel pain—indeed, you could have the full range of sensory experience.